Install PyTorch on Apple M1 (M1, Pro, Max) with GPU (Metal)

Note:

  • Uninstall Anaconda/Anaconda Navigator and other related previously installed version of conda-based installations.
  • Anaconda and Miniforge cannot co-exist together.
  • However, if you have previous installations of PyTorch with Miniforge, you can continue to use that without uninstalling it.
  • We will install the GPU version in a new conda environment anyway.
  • Keep in mind that the GPU version is still a nightly build and expect it to have breaking changes for the next couple of months.

Installing Miniforge

  • Install miniforge from brew: brew install miniforge
  • Create an anaconda environment: conda create -n pt122
  • Activate the environment: conda activate pt122
  • Install Python: conda install python

Installing PyTorch

  • Run: pip install --pre torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/nightly/cpu to install PyTorch (Nightly) dependencies
  • Execute: pip install pandas jupyterlab to install relevant dependencies to run sample code.
  • Run: conda activate pt122

Verifying Installation

  • Execute: jupyter-lab to open a Jupyter Notebook and run the following code:
import torch
torch.device("mps")
torch.__version__
torch.tensor([1,2,3], device="mps")

Checking GPU

Credits